Background: Systemic Lupus Erythematosus (SLE) is a highly heterogeneous systemic autoimmune disease characterized by complex clinical manifestations and molecular dysregulation. Although two biologic therapies targeting B-cell differentiation and type I interferon pathways are currently approved for SLE, several others are used off-label. However, there is no evidence-based molecular criteria to guide biologic selection, leading to trial-and-error treatment strategies that may compromise patient outcomes. Previous studies suggest that systemic autoimmune diseases (SADs) share pathogenic pathways and that molecular signatures may support more precise therapeutic decision-making.

Objectives: To evaluate the differential predicted responses to biologic therapies using pathway-based molecular stratification in order to improve treatment selection beyond clinical diagnosis alone and achieve a more personalized medicine.

Methods: Unsupervised factorization with MOFA was applied to PRECISESADS transcriptomic data from over 1,500 individuals across seven systemic autoimmune diseases. Molecular factors were characterized using Gene Set Enrichment Analysis to identify immune pathway associations. These pathway-associated factors were used for a M3C consensus clustering to define robust molecular clusters (pathotypes). In-silico drug response simulations were performed with hipathia based on molecular pathways targeted by commonly used biologics, including BAFF, IFNAR, TNF, CD20, and IL1R1. Pathotype stratification was subsequently validated with an independent SLE cohort (GSE88887).

Results: Molecular pathotypes demonstrated higher in-silico response rates for specific biologic targets compared with clinical diagnosis alone. For example, BAFF inhibition showed high predicted response rates within interferon-associated and plasma cell-associated pathotypes. Clinical phenotypes were distributed across all molecular pathotypes, with each phenotype displaying response trends consistent with overall pathotype patterns. These findings support shared biological mechanisms across SADs that can be more effectively treated with biologics than using traditional therapeutic approaches. Pathotype classification and predicted response rates were successfully validated in the independent SLE cohort.

Conclusions: Pathway-based molecular stratification identifies biologically distinct patient subgroups with differential predicted responses to biologic therapies, improving the responses when using clinical diagnosis alone. The consistent distribution of clinical phenotypes across molecular pathotypes supports a molecular pathway-driven therapeutic approach across systemic autoimmune diseases. Validation in an independent cohort demonstrates the robustness and clinical potential of this strategy, supporting the future design of molecularly guided precision medicine trials in SLE and related diseases.

REFERENCES: NIL.

Acknowledgments: NIL.

Disclosure of Interests: None declared.